Give every line a spec, a destination and an approver. Video projects rarely overrun on the film. They overrun on the eleven versions of the film that nobody itemised.

Category
Filmmaking
Author

Justkay
Documentary Filmmaker & Founder at Storyflow
Topics
2026-08-10
•
13 min read
•
FilmmakingTable of Contents
Build a deliverables list by giving every item three things: a specification, a destination, and an approver. A line that says "social cuts" is not a deliverable, it is an argument you will have in week six. A line that says "one 9x16 cut, 30 seconds, burned-in captions, for Instagram Reels, approved by Priya" is a deliverable, because there is exactly one thing it could mean and exactly one person who can say it is finished. Most video projects that overrun do not overrun on the film. They overrun on the eleven versions of the film that nobody itemised. A deliverable nobody named a destination for is a deliverable you will make twice.
In week six somebody asks why there are two vertical cuts, and the reason was agreed in a call nobody recorded. Storyflow keeps the brief, the references and the decisions on one canvas the AI reads end to end. Paid-only during early access; the Free plan lands before the end of 2026.

Take any line on a deliverables list and check it against three questions. If you cannot answer all three, it is not finished.
The specification is what the file is. Aspect ratio, duration, resolution, frame rate, audio configuration, whether captions are burned in or supplied separately, whether there is a version with no text for later localisation. Precision here is not pedantry; each of these is a different export and some are different edits.
The destination is where it will play. Not "social", which is four platforms with different requirements, but Instagram Reels, or a LinkedIn feed post, or a trade-show loop, or the homepage hero. Destination is what determines spec, which is why it comes first in the conversation and last in most lists.
The approver is who says it is done. Usually the same person for everything, and occasionally not: a regional lead may own the localised versions, legal may own anything with a claim in it. Naming the exception at the start prevents the week where a finished file waits on someone nobody thought to include.
The check that catches most problems: read the line to yourself and ask whether two reasonable people could produce different files from it. If yes, the line is not specific enough. "A short version" fails. "One 30 second cut, 16x9, no captions, for the pre-roll placement" passes.
The single most common failure is the uncounted plural.
"Social cuts", "some stills", "a few teasers", "language versions". Each of these is a number nobody has agreed on, and in the client's head the number is larger than in yours. It costs nothing to fix at the start and a weekend to fix at the end.
Write a number on every plural. Three social cuts, not social cuts. Ten stills, not a selection. Two language versions, not localisation.
When the number genuinely is not known yet, price the unit. "Additional 9x16 cuts at [amount] each" converts an unknown into a menu, and it means the answer to "can we get three more?" is a cheerful yes rather than a difficult conversation.
Watch for the plural hiding inside a singular. "A film" often means a film plus a trailer plus a teaser plus a cutdown, because that is what the last agency delivered. Ask directly whether anything shorter is expected alongside the main piece.
These are real work and they are almost never on the first draft of a list.
Captions and subtitles. A burned-in caption version is a separate export and often a separate design pass. An SRT or VTT file is a separate deliverable again. Platforms increasingly need one or both, and neither is free.
Thumbnails and cover frames. A YouTube thumbnail is a design job, not a screenshot. A Reels cover frame is a different crop. Both get requested at the last minute because they are invisible until upload.
Stills. Frame grabs are cheap; usable stills often are not, because the frame that looks right in motion is rarely the frame that looks right paused.
A textless or clean version. Any project that might be localised or reused later wants a master with no burned-in text. Producing it after the fact means reopening the project; producing it during delivery costs minutes.
The archive. A versioned, named archive of the project and its assets, handed over or stored. If the client expects to come back in a year, this is a deliverable, and if nobody names it, it is a favour you do badly under time pressure.
Music and stock licences. Not a file, but a deliverable in the sense that someone has to receive proof of what is licensed and for how long. This is the one that becomes a legal problem rather than a scheduling one.
The reason to ask where before you agree what is that platforms impose the answer.
A 30 second cut for a paid placement has a hard duration limit and often a specification sheet from the platform. A homepage hero probably wants no audio by default and a loop point. A trade-show screen may be an unusual ratio and will be viewed from six metres away, which changes type size. A LinkedIn feed post is watched muted by default, which makes captions mandatory rather than nice.
Ask for the placement, not the platform, when money is involved. Paid placements have specification documents. Getting that document in week one is much better than reformatting in week six because the ad platform rejected the file.
When the destination is genuinely unknown, deliver the most flexible master and price the adaptations separately. A clean 16x9 master at full resolution adapts to almost anything; a burned-in, music-locked, 30 second vertical cut adapts to nothing.
A deliverables list that lists only deliverables leaves everything else ambiguous, and ambiguity resolves toward more work.
Add a short block underneath. Not included: additional aspect ratios beyond those listed, additional language versions, subtitle files beyond the listed formats, thumbnail design, additional revision rounds beyond two, ongoing hosting, and raw footage.
Raw footage deserves its own mention because it is the most commonly assumed deliverable in video. Whether the client receives the rushes is a real commercial question with storage implications, and it should be answered explicitly at kickoff rather than discovered at the end.
Price the exclusions where you can. A list with prices next to the excluded items is a menu, and a menu produces sales rather than disputes.
For a two minute brand film with a social rollout, the list a client can actually approve:
| Deliverable | Spec | Destination | Approver |
|---|---|---|---|
Hero film | 2:00, 16x9, 1080p, stereo, no burned-in captions | Homepage and YouTube | Priya |
Hero film, captioned | 2:00, 16x9, burned-in captions | LinkedIn feed | Priya |
Cutdown | 0:30, 16x9, no captions | Paid pre-roll, spec sheet to follow | Priya and legal |
Vertical cut A | 0:30, 9x16, burned-in captions | Instagram Reels | Priya |
Vertical cut B | 0:15, 9x16, burned-in captions | TikTok | Priya |
Subtitle files | SRT for the hero film, English | Uploaded alongside | Priya |
Stills | 10 selected and lightly graded | Press and social | Marco |
Textless master | 2:00, 16x9, no text or captions | Held for future localisation | Priya |
Project archive | Named, versioned, delivered on drive | Client storage | Marco |
Nine lines, all countable, all with a home. The version of this list that causes trouble reads: hero film, social cuts, some stills, captions.
Note that the cutdown has two approvers, because it carries a product claim. That is the kind of detail that costs a week if it is discovered at delivery and costs nothing if it is written at kickoff.
The list is a table. A spreadsheet, a page in your project tool, or a table in the statement of work all work, and the format matters far less than whether the client approved it in writing.
If you use a project management tool, make each deliverable a task with the spec in the description and the approver as a watcher, so status is visible without asking. Asana, Monday, ClickUp and Notion all do this adequately.
The list and the approval are two different artifacts, and most projects only keep the first one. The table says a 9x16 cut with burned-in captions was scoped. It does not say Priya approved version three of it on 14 August. When a dispute arrives, the second document is the one that ends it, and it usually does not exist because approval happened in a reply to an email thread that has since been buried.
| What | What it has to survive | Where it goes |
|---|---|---|
The list itself | Being agreed once and referenced for months | A spreadsheet, or a table in the statement of work |
The status of each item | Daily questions about what is done | A project tool, one task per deliverable |
The approval of each item | A dispute six weeks later | The platform the client approved in, with a version and a date |
For the third row, a review platform that carries an approval state per version does the recording for you. Frame.io is the standard when an editor works in Premiere or After Effects, because its panels put the notes where the work happens. Framekit is the flat-priced alternative when the reviewer is the client rather than a post house: comments land on a timecode, versions stack with exactly one current, and each version holds its own status of needs review, in progress, changes requested, or approved. Either way, the point is that "approved" becomes a field attached to a specific version rather than a sentence in an inbox.
Delivery is worth separating from review, too. The review link is iterative and should expire. The delivery link is final and needs download control, because a master file and a social cut have different audiences inside the client's company. Handing over a shared-drive folder collapses both into one permission and leaves no record of what was actually sent. A delivery gallery or page with viewing and downloading as separate toggles fixes that, and takes about a minute to set up per project.
Storyflow is our product and it is not the tool for this. It has no task management, no status tracking, no assignees and no approval workflow, so the list itself belongs in a spreadsheet or a project tool. The adjacent job it does is upstream: holding the brief, the references and the decisions on one canvas so the reason a deliverable exists is recoverable when someone asks in week six why there are two vertical cuts. Paid-only during early access, with the Free plan landing before the end of 2026, and anyone a paid member invites to a board joins free now.
Framekit is also ours, and it is honest to say it does not build the list either. It has no task management and no assignees. It covers the two rows underneath: the approval record and the delivery. Free includes video review with 2GB of video storage and unlimited client galleries with 3GB of photo storage; paid tiers run $9 to $39 per month, flat per account.
Deliverables lists fail quietly. Nobody argues about them at the start, which is exactly why they are worth twenty minutes at the start.
Give every line a spec, a destination and an approver. If a line is missing one, that is where the week six conversation will happen.
Put a number on every plural, and price the unit when the number is genuinely unknown.
Scope the invisible items: captions in both forms, thumbnails, stills, a textless master, the archive and the licences.
Ask where before you agree what, because the destination writes the specification.
Then put the exclusions next to the list, priced, so the extras are a menu rather than a dispute.
A deliverable nobody named a destination for is a deliverable you will make twice.
If two reasonable people could produce different files from the line, it is not specific enough. Each line needs a spec, a destination and an approver, and every plural needs a number.
Only if it is explicitly agreed, and it should be discussed at kickoff. It is the most commonly assumed deliverable in video, it has real storage and delivery costs, and it changes what you can reuse later. Put it in the exclusions if it is not included.
Price the unit. "Additional 9x16 cuts at [amount] each" turns an unknown quantity into a menu, so more cuts become a cheerful transaction instead of a scope dispute.
Yes, and often two: a burned-in version, which is a separate export and sometimes a separate design pass, and a subtitle file such as SRT or VTT. Neither is included by default in most people's mental model of "the film".
A version with no burned-in text or captions. You need it if there is any chance of localisation or reuse. Producing it during delivery costs minutes; producing it a year later means reopening an archived project.
At kickoff. At kickoff it is a menu the client chooses from; at delivery it is a negotiation you will lose. It also drives the estimate, so it has to exist before the number does.
Usually one person for everything, with exceptions worth naming explicitly: legal for anything carrying a claim, a regional lead for localised versions, a brand owner for anything using the logo unusually. Naming exceptions at the start prevents finished files waiting on unknown reviewers.
There is no normal, which is exactly why the number has to be agreed rather than assumed. What matters is that the list says a number and a duration and a platform for each one.
A design job, and a separate deliverable. It gets requested at upload time, which is the worst possible moment, and it is one of the most commonly forgotten lines.
It should be in whatever the client formally approves, which for most projects is the statement of work or the estimate. What matters is written approval, not the document's name.
The list drives the estimate. Delivery is consistently underestimated because it feels like a formality, and an itemised list is what turns it into countable work.
Asking where each item will play before agreeing what it is. Destination determines specification, and reversing that order is how you end up remaking files to meet a platform requirement nobody mentioned.
Against a specific version, in the platform the client used, with a date. An email saying "looks good" is a memory rather than a record, because it does not say which cut it referred to and it is unfindable in six weeks. Review platforms that carry an approval status per version (Frame.io, Framekit, and others) do the recording as a side effect of the reviewer clicking a button, which is why that beats any convention you have to remember to follow.
No, and conflating them is a small but reliable source of trouble. A review link is iterative, should expire, and usually should not allow downloads, because a rough cut circulating internally gets shown in meetings you are not in. A delivery link is final and needs download control per audience. Some tools cover both on one account, which is fine, but keep them as separate links with separate permissions.
No. The list defines what is being made; the contract defines what happens when that changes, who owns the work, and when you get paid. A good statement of work contains the list, which is the cleanest arrangement, because then a scope change is visibly a change to a document both parties signed rather than an argument about what was implied.
Every Storyflow board starts from real structure and an AI that reads the whole canvas. Open one of these templates and make it yours.
A visual AI workspace where every feature lives inside one canvas. No tab-switching, no context lost.
Build your entire board from a single message
Type what you need in the AI chat at the bottom of your canvas. The AI adds cards, headings, and structure directly onto your board.
Use expert frameworks as AI context
Type @ in the AI chat and choose any Tactic. The AI tailors every response to that framework instead of giving generic advice.
Turn your board into a mind map in seconds
Ask the AI to restructure your canvas as a mindmap. It connects your ideas into a visual hierarchy so you can see how everything relates.
Storyflow actually began as a personal tool while working on creative and research projects.
We kept running into the same problem: ideas were scattered everywhere: notes, documents, and whiteboards.
Nothing helped us see how everything connected.
So we started building a workspace designed around how ideas actually grow.
→ Read how Storyflow was created
Justkay
Documentary Filmmaker & Founder at Storyflow
Published: 2026-08-10
Transform your creative workflow with AI-powered tools. Generate ideas, create content, and boost your productivity in minutes instead of hours.